LEVERAGING DATA FOR
A TRUE OMNI-CHANNEL
EXPERIENCE
NIREN SIROHI, PH.D.
PA R T N E R , G L O B A L D ATA S C I E N C E L E A D , I K N O W T I O N
A TELETECH COMPANY
Niren Sirohi, Ph.D.
Partner, Global Data Science Lead
Leveraging Data for a
True Omni-Channel Experience
3
Ph.D. in Management at Cornell's Johnson Graduate School of Management, MBA
with a major in marketing, and graduate degree in Mechanical Engineering
Today’s Presenter
■ Develop & implement strategic analytic
solutions
■ Global brands -- financial, auto, retail,
telecomm, hospitality
■ Hobbies: technology & rock concerts
4
Apply Analytic and
Consultative Best Practices to:
 Identify customer-level
opportunities
 Inform customer-centric
strategies
 Maximize customer-focused
programs
 Measure customer-driven
impacts
iKnowtion – Leaders in Analytics
5
Agenda for Today
■ VoC and why it’s important
■ CX strategy and customer behavior
■ True omnichannel experience
■ Key ingredients for success
■ Getting started
6
Source: The State of Marketing 2013, IBM’s Global Survey of Marketers
$83 billion lost each year
7
State of VoC today
80% of companies say they
deliver “superior” customer
service. 8% of people think
these same companies
deliver “superior” customer
service
A typical
business only
hears from 4%
of its dissatisfied
customers
Only 2-7% of companies
self identified as being
very good across
various VoC
competencies
70% of companies are
collecting customer
feedback, only one-third
actually believe that they
are designing solutions to
problems based on the
insights
Only 11% of companies have
reached the top 2 stages of
VoC maturity (collaborators
and transformers)
Source: Temkin Group and Customer Service Facts, Quotes and Statistics
8
VoC Myths
VoC = Explicit Feedback
VoC Data is Sufficient
VoC Only from Customers
VoC = Survey Data
9
Your customers are sharing
their needs
Please press 1
for service, 2
for billing…
10
But nobody’s
connecting the dots…
11
Information is siloed by channel and within enterprises.
CUSTOMERINFORMATION
CUSTOMERINFORMATION
CUSTOMERINFORMATION
CUSTOMERINFORMATION
CUSTOMERINFORMATION
CUSTOMERINFORMATION
CUSTOMERINFORMATION
EDWCRM
12
Information isn’t
transformed into insights
and next best actions
Need predictive and
prescriptive analytics
13
Insights and next best actions not relevant
and inconsistently shared across the firm
14
Information is unable to
be used in the moment
15
Everyone in the organization can benefit from VoC
Product
Management
Pricing
CRM
Customer
Care
Customer
Experience
Sales
Marketing and
PR
Early, continuous, real-time input on customer perceptions to
pricing strategy, competitive pricing, and gaps
Identify next best action for individual customers and prospects
based on contextual, in the moment data and historical behavior
Drive efficiency and effectiveness of care operations by enabling
agents with timely alerts and recommendations. Identify sales and
CRM leads
Understand key pain points and influence the customer journey to
reduce customer effort, reduce detractors, and build advocates
Gain insight into competitive sales strategy, prospect needs,
perceptions, and decision factors
Early warning indicators on brand health and perceptionsEarly, continuous, real-time input on product features,
functionality, performance, competitive differentiators
16
Signals from VoC can lead to action at the aggregate level…
“I cannot reset my password
online.”
“The web portal would not let
me add this feature.”
“Why can’t I check /
reschedule my appointment
online?”
“I can’t see the current
status of my install order
online.”
“Company X gave me a
great deal on triple play
when I came off promo
pricing at brand Y.”
“They are fleecing you with
off-promo price increases.”
+ CRM
data
+ CRM
data
+ CRM
data
High value customers are
not able to complete
account management tasks
online and have to call in
Customers in the first 30
days are frustrated to not be
able to check status and
schedule appts online
Competitors are wooing our
most valuable customers
away when they come off
promo period
Fix the web portal
and reduce call
volume
Build an
appointment and
order management
application and
reduce call volume
Understand price
elasticity and
adjust off-promo
pricing strategy to
maximize retention
VoC Signal Action
Improve Operational Efficiencies (reduce call volume) & Improve Customer Satisfaction
Inform business strategy and customer lifecycle management
17

…as well as at the individual level
Execute
trades
Positive
comment on
Facebook
Adds money
to brokerage
account
VoC Platform
CRM
Web
Interactions
Jack placed a large sell
order today that did not
execute
Yesterday, Jack made a
negative comment about
our fund platform on our
Facebook page
A week ago, high value
customer Jack tried to
watch a video on our
website that failed to load
Real Time
Listening
Jack is a high value, high
risk customer who should
get a call from his advisor:
■ Acknowledge issues
■ Offer free trades
■ Ask for recommendation
if satisfied
ALERT!
Real Time
Decisions
Wealth
Management
Advisor
Phone Call
..... ....
Story begins…
...Happy Ending



3:50 pm
4:00 pm
18
Benefit Examples
Insights Drive Customer Experience Improvements:
■ Optimizes interactions throughout the customer
journey
■ Pinpoints CX improvement opportunities with the
highest expected ROI
Intelligence Leads to Greater Customer Value:
■ Helps convert brand detractors into brand loyalists
and advocates
■ Guides the development of expanded value
propositions, leading to higher share of wallet
Customer Feedback Fuels Innovation:
■ Highlights existing product enhancements and new
product opportunities
■ Tracks brand perceptions and drives customer-
centric business transformations
75% Improvement in
repeat call prediction
3X Uplift
NPS Scores
Increase in
Customer
Retention
15%
Higher
Customer
Value20%
19
Key Ingredients for Success
Real time listening across touch-points
Data integration across touch-points and enterprise systems
Disseminate and share signals with relevant stakeholders and
give them actionable recommendations
Integrate with frontline systems like CRM, contact centers etc.
Efficient Signal Extraction from structured and
unstructured data
20
Unstructured Data Explains “Why”
Structured
(Quantitative)
Unstructured
(Qualitative)
Customer-Initiated Company-Initiated
Voice of Customer Categories
DepthofInsight
Representativeness of Feedback
21
Analysis of text, speech, & contextual data across channels
Text
Dear DirectCom Online Service, I have a question about my
most recent bill. I paid the full balance online in the amount of
47.61$ on February 16th. When I checked back the payment
didn’t show up, and I was charged 50$ fee on top of that. This
is not acceptable! The online payment confirmation number
back the payment didn’t show up, and I was charged 50 and I
was charged 50 and I was $ fee on top of that. This is not is
49733. Please verify and make sure I get the fee back to my
account. This is not the first time I have a problem with the
online payment. Best regards, Gina Lowell.
SpeechContext
Misinformed
Frustrating
This is the 2nd time I’m calling
Phone
Social
Media
Chat
Email
Web
Phone
Survey
Keywords and Topics
Talk Over Analysis
Emotion Detection
Call Flow Analysis
So confusing
Sick and tired
Desktop Analytics
Customer Demographics
Interaction History
Customer Feedback
Keywords and Topics
Sentiment Analysis
Chat Response Time
Social Buzz
22
Customer
Service
Knowledge /
Helpfulness
Attitude
Multiple
Attempts
Tier 2 Escalation
Problem
Resolution
Troubleshooting
Follow up
Quality of
Service
Refunds /
Discounts
Customer
Support Rep
Communication
Skills
Language
Barrier
Listening
Speed of service
Accessibility /
Reaching Rep
Transfer / Hold
Reaching Right
Person
IVR
Business Hours
Phone
Experience
Runaround
Hold Time
Transfers
Real Person
… … … …
Branches
(19)
Leaves
(111)
Illustrative
Domain knowledge and Machine Learning transforms
unstructured data into structured insights
…
23
Insights help improve contact center operations
Attitude of reps is a clear strength
Troubleshooting abilities, communication and
language skills need improvement
Improve routing skills and workforce
management to minimize hold times and get
customers to the right rep the first time
Telco Call Center Speech and Text Analytics
Category Leaves
Volume,Sentiment
24
Incremental NPS Changes by Making Driver Improvements
CX improvements lead to significant improvement in NPS
Automotive CX Drivers
-29
19
17
15
13
10
11
7
6
1
70
-40
-20
0
20
40
60
80
Problem
Resolution
Beginning
NPS
Hold Time Personal
Effort
Providing Best
Help
Dealer
Resolution
Attempt
Listening Courtesy Automated
Phone Ease of
Use
# of Consumer
Affair contacts
Final NPS
Structured
Unstructured
25
Insights can be incorporated into Predictive Models
1 2 3 4 5 6 7 8 9 10
ModelPerformance
Structured plus unstructured Structured only
Observed Benefit by Model Deciles
Variation
explained by
structured
data only
Variation
explained by
both
structured
and
unstructured
data
Illustrative
Unstructured data helps explain more of the
observed variation when developing predictive
customer behavioral models
26
Develop Strategy
and Prove the Value
First 3-4 months
■ Assess current VOC
data being collected
and efforts to derive
insight and act upon
them
■ Develop VOC
roadmap and strategy
■ Integrate and analyze
VOC data on a pilot
basis to build
business case
Build Capabilities and
Implement
Next 5-6 months
■ Collect requirements
from different
stakeholders who
would value from
VOC insights
■ Develop capabilities
(people, process and
infrastructure) to
collect, analyze,
share, respond and
act on VOC data
Accelerate Impact
On-Going
■ Identify test and learn
opportunities
■ Launch, measure and
refine VOC programs
■ Add additional data
sources and
stakeholders
■ Refine capabilities
An omnichannel experience requires a multi-step journey
1 2 3
27
■ Commit to continuously listening to and
analyzing the Voice of your Customer (VoC)
■ Increase focus on predictive and
prescriptive analytics
■ Focus on customer journeys, not
interactions
■ Detect meaningful signals in VoC and act
on them
Summary
28
Niren Sirohi, Ph.D.
Partner, Global Data Science Lead
nsirohi@iKnowtion.com
(781) 494.9989 x201
A TELETECH COMPANY
Thank You
QUESTIONS?

Leveraging Data for a True Omnichannel Experience

  • 1.
    LEVERAGING DATA FOR ATRUE OMNI-CHANNEL EXPERIENCE NIREN SIROHI, PH.D. PA R T N E R , G L O B A L D ATA S C I E N C E L E A D , I K N O W T I O N A TELETECH COMPANY
  • 2.
    Niren Sirohi, Ph.D. Partner,Global Data Science Lead Leveraging Data for a True Omni-Channel Experience
  • 3.
    3 Ph.D. in Managementat Cornell's Johnson Graduate School of Management, MBA with a major in marketing, and graduate degree in Mechanical Engineering Today’s Presenter ■ Develop & implement strategic analytic solutions ■ Global brands -- financial, auto, retail, telecomm, hospitality ■ Hobbies: technology & rock concerts
  • 4.
    4 Apply Analytic and ConsultativeBest Practices to:  Identify customer-level opportunities  Inform customer-centric strategies  Maximize customer-focused programs  Measure customer-driven impacts iKnowtion – Leaders in Analytics
  • 5.
    5 Agenda for Today ■VoC and why it’s important ■ CX strategy and customer behavior ■ True omnichannel experience ■ Key ingredients for success ■ Getting started
  • 6.
    6 Source: The Stateof Marketing 2013, IBM’s Global Survey of Marketers $83 billion lost each year
  • 7.
    7 State of VoCtoday 80% of companies say they deliver “superior” customer service. 8% of people think these same companies deliver “superior” customer service A typical business only hears from 4% of its dissatisfied customers Only 2-7% of companies self identified as being very good across various VoC competencies 70% of companies are collecting customer feedback, only one-third actually believe that they are designing solutions to problems based on the insights Only 11% of companies have reached the top 2 stages of VoC maturity (collaborators and transformers) Source: Temkin Group and Customer Service Facts, Quotes and Statistics
  • 8.
    8 VoC Myths VoC =Explicit Feedback VoC Data is Sufficient VoC Only from Customers VoC = Survey Data
  • 9.
    9 Your customers aresharing their needs Please press 1 for service, 2 for billing…
  • 10.
  • 11.
    11 Information is siloedby channel and within enterprises. CUSTOMERINFORMATION CUSTOMERINFORMATION CUSTOMERINFORMATION CUSTOMERINFORMATION CUSTOMERINFORMATION CUSTOMERINFORMATION CUSTOMERINFORMATION EDWCRM
  • 12.
    12 Information isn’t transformed intoinsights and next best actions Need predictive and prescriptive analytics
  • 13.
    13 Insights and nextbest actions not relevant and inconsistently shared across the firm
  • 14.
    14 Information is unableto be used in the moment
  • 15.
    15 Everyone in theorganization can benefit from VoC Product Management Pricing CRM Customer Care Customer Experience Sales Marketing and PR Early, continuous, real-time input on customer perceptions to pricing strategy, competitive pricing, and gaps Identify next best action for individual customers and prospects based on contextual, in the moment data and historical behavior Drive efficiency and effectiveness of care operations by enabling agents with timely alerts and recommendations. Identify sales and CRM leads Understand key pain points and influence the customer journey to reduce customer effort, reduce detractors, and build advocates Gain insight into competitive sales strategy, prospect needs, perceptions, and decision factors Early warning indicators on brand health and perceptionsEarly, continuous, real-time input on product features, functionality, performance, competitive differentiators
  • 16.
    16 Signals from VoCcan lead to action at the aggregate level… “I cannot reset my password online.” “The web portal would not let me add this feature.” “Why can’t I check / reschedule my appointment online?” “I can’t see the current status of my install order online.” “Company X gave me a great deal on triple play when I came off promo pricing at brand Y.” “They are fleecing you with off-promo price increases.” + CRM data + CRM data + CRM data High value customers are not able to complete account management tasks online and have to call in Customers in the first 30 days are frustrated to not be able to check status and schedule appts online Competitors are wooing our most valuable customers away when they come off promo period Fix the web portal and reduce call volume Build an appointment and order management application and reduce call volume Understand price elasticity and adjust off-promo pricing strategy to maximize retention VoC Signal Action Improve Operational Efficiencies (reduce call volume) & Improve Customer Satisfaction Inform business strategy and customer lifecycle management
  • 17.
    17  …as well asat the individual level Execute trades Positive comment on Facebook Adds money to brokerage account VoC Platform CRM Web Interactions Jack placed a large sell order today that did not execute Yesterday, Jack made a negative comment about our fund platform on our Facebook page A week ago, high value customer Jack tried to watch a video on our website that failed to load Real Time Listening Jack is a high value, high risk customer who should get a call from his advisor: ■ Acknowledge issues ■ Offer free trades ■ Ask for recommendation if satisfied ALERT! Real Time Decisions Wealth Management Advisor Phone Call ..... .... Story begins… ...Happy Ending    3:50 pm 4:00 pm
  • 18.
    18 Benefit Examples Insights DriveCustomer Experience Improvements: ■ Optimizes interactions throughout the customer journey ■ Pinpoints CX improvement opportunities with the highest expected ROI Intelligence Leads to Greater Customer Value: ■ Helps convert brand detractors into brand loyalists and advocates ■ Guides the development of expanded value propositions, leading to higher share of wallet Customer Feedback Fuels Innovation: ■ Highlights existing product enhancements and new product opportunities ■ Tracks brand perceptions and drives customer- centric business transformations 75% Improvement in repeat call prediction 3X Uplift NPS Scores Increase in Customer Retention 15% Higher Customer Value20%
  • 19.
    19 Key Ingredients forSuccess Real time listening across touch-points Data integration across touch-points and enterprise systems Disseminate and share signals with relevant stakeholders and give them actionable recommendations Integrate with frontline systems like CRM, contact centers etc. Efficient Signal Extraction from structured and unstructured data
  • 20.
    20 Unstructured Data Explains“Why” Structured (Quantitative) Unstructured (Qualitative) Customer-Initiated Company-Initiated Voice of Customer Categories DepthofInsight Representativeness of Feedback
  • 21.
    21 Analysis of text,speech, & contextual data across channels Text Dear DirectCom Online Service, I have a question about my most recent bill. I paid the full balance online in the amount of 47.61$ on February 16th. When I checked back the payment didn’t show up, and I was charged 50$ fee on top of that. This is not acceptable! The online payment confirmation number back the payment didn’t show up, and I was charged 50 and I was charged 50 and I was $ fee on top of that. This is not is 49733. Please verify and make sure I get the fee back to my account. This is not the first time I have a problem with the online payment. Best regards, Gina Lowell. SpeechContext Misinformed Frustrating This is the 2nd time I’m calling Phone Social Media Chat Email Web Phone Survey Keywords and Topics Talk Over Analysis Emotion Detection Call Flow Analysis So confusing Sick and tired Desktop Analytics Customer Demographics Interaction History Customer Feedback Keywords and Topics Sentiment Analysis Chat Response Time Social Buzz
  • 22.
    22 Customer Service Knowledge / Helpfulness Attitude Multiple Attempts Tier 2Escalation Problem Resolution Troubleshooting Follow up Quality of Service Refunds / Discounts Customer Support Rep Communication Skills Language Barrier Listening Speed of service Accessibility / Reaching Rep Transfer / Hold Reaching Right Person IVR Business Hours Phone Experience Runaround Hold Time Transfers Real Person … … … … Branches (19) Leaves (111) Illustrative Domain knowledge and Machine Learning transforms unstructured data into structured insights …
  • 23.
    23 Insights help improvecontact center operations Attitude of reps is a clear strength Troubleshooting abilities, communication and language skills need improvement Improve routing skills and workforce management to minimize hold times and get customers to the right rep the first time Telco Call Center Speech and Text Analytics Category Leaves Volume,Sentiment
  • 24.
    24 Incremental NPS Changesby Making Driver Improvements CX improvements lead to significant improvement in NPS Automotive CX Drivers -29 19 17 15 13 10 11 7 6 1 70 -40 -20 0 20 40 60 80 Problem Resolution Beginning NPS Hold Time Personal Effort Providing Best Help Dealer Resolution Attempt Listening Courtesy Automated Phone Ease of Use # of Consumer Affair contacts Final NPS Structured Unstructured
  • 25.
    25 Insights can beincorporated into Predictive Models 1 2 3 4 5 6 7 8 9 10 ModelPerformance Structured plus unstructured Structured only Observed Benefit by Model Deciles Variation explained by structured data only Variation explained by both structured and unstructured data Illustrative Unstructured data helps explain more of the observed variation when developing predictive customer behavioral models
  • 26.
    26 Develop Strategy and Provethe Value First 3-4 months ■ Assess current VOC data being collected and efforts to derive insight and act upon them ■ Develop VOC roadmap and strategy ■ Integrate and analyze VOC data on a pilot basis to build business case Build Capabilities and Implement Next 5-6 months ■ Collect requirements from different stakeholders who would value from VOC insights ■ Develop capabilities (people, process and infrastructure) to collect, analyze, share, respond and act on VOC data Accelerate Impact On-Going ■ Identify test and learn opportunities ■ Launch, measure and refine VOC programs ■ Add additional data sources and stakeholders ■ Refine capabilities An omnichannel experience requires a multi-step journey 1 2 3
  • 27.
    27 ■ Commit tocontinuously listening to and analyzing the Voice of your Customer (VoC) ■ Increase focus on predictive and prescriptive analytics ■ Focus on customer journeys, not interactions ■ Detect meaningful signals in VoC and act on them Summary
  • 28.
    28 Niren Sirohi, Ph.D. Partner,Global Data Science Lead nsirohi@iKnowtion.com (781) 494.9989 x201 A TELETECH COMPANY Thank You
  • 29.